Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Flats flats flats (again)

  • Please log in to reply
13 replies to this topic

#1 limeyx

limeyx

    Ranger 4

  • -----
  • topic starter
  • Posts: 355
  • Joined: 23 May 2020
  • Loc: Seattle, WA

Posted 25 September 2020 - 10:10 PM

So I have had a big fight with these and there are still some baffling things. I wanted a scientific way to take and evaluate them but I don't have $250+ for a "real" flat field generator

 

I got a LED tracing box from Amazon and set it to the brightest setting with three white sheets of copy paper on top (after much messing around) and I finally (think) I have a good flat ... but there is still confusing data

 

Link to flat

 

https://drive.google...iew?usp=sharing

I gauged this with "Iris", setting its histogram to max 16384

 

This is what Iris tells me (looks like I am clumped around 9500 ADU which is a little over half but seems reasonable)

 

https://drive.google...iew?usp=sharing

 

However, the APT histogram is ... way off to the right

 

https://drive.google...iew?usp=sharing

 

And GIMP doesn't really want to display a histogram

 

This is ISO 200, 1/8sec at f/4.5

 

I know that ideally flats ISO should match lights, and currently I plan to shoot lights at 200 as per other recommendations on the forums

 

It will be frustrating if I end up needing to use different ISO's for lights because I don't have good control over the panel brightness and I don't want to have too fast a shutter speed since it causes other issues...May need more paper sheets ?

 

 



#2 SilverLitz

SilverLitz

    Apollo

  • -----
  • Posts: 1,160
  • Joined: 17 Feb 2018
  • Loc: Louisville, KY

Posted 25 September 2020 - 10:23 PM

Use APT's Flats Aid tool and set ADU target for 20,000.  This will calc the flat exposure.


  • limeyx likes this

#3 limeyx

limeyx

    Ranger 4

  • -----
  • topic starter
  • Posts: 355
  • Joined: 23 May 2020
  • Loc: Seattle, WA

Posted 25 September 2020 - 10:38 PM

Use APT's Flats Aid tool and set ADU target for 20,000.  This will calc the flat exposure.

Isn't that only available for a CCD camera ? I am using a DSLR and I dont see that it works on DSLR unless I am missing something

Also, unfortunately my D5300 doesn't have the pins needed to allow Av mode with my lens so I only have manual mode available



#4 russellmm

russellmm

    Vostok 1

  • -----
  • Posts: 151
  • Joined: 02 Aug 2020
  • Loc: Gig Harbor WA

Posted 25 September 2020 - 10:47 PM

yea, APT wont do it, but Sequencer Generator Pro will do the same thing and works with DSLRs. That is what I ended up using.


Edited by russellmm, 25 September 2020 - 10:47 PM.


#5 Dynan

Dynan

    Skylab

  • *****
  • Posts: 4,396
  • Joined: 11 Mar 2018
  • Loc: NOLA

Posted 25 September 2020 - 10:48 PM

The flat looks really good. Maybe a little underexposed...half scale of ≈ 7K would be great. But you're not touching either edge of the histo, so you should be good. Here's what PI said about it:

 

FLAT 1.jpg

 

Your flat contour plot is the cleanest I've ever seen as far as vignetting and lack of dust donuts.waytogo.gif


  • limeyx likes this

#6 limeyx

limeyx

    Ranger 4

  • -----
  • topic starter
  • Posts: 355
  • Joined: 23 May 2020
  • Loc: Seattle, WA

Posted 25 September 2020 - 10:57 PM

The flat looks really good. Maybe a little underexposed...half scale of ≈ 7K would be great. But you're not touching either edge of the histo, so you should be good. Here's what PI said about it:

 

attachicon.gifFLAT 1.jpg

 

Your flat contour plot is the cleanest I've ever seen as far as vignetting and lack of dust donuts.waytogo.gif

Thanks for looking. The camera is used but I got it with <1000 shutter clicks EDIT: This may explain the lack of dust

 

Also the lens is an old Manual Focus Nikon 300 full frame lens so it makes sense it should look pretty good on APS-C

Its just so weird that APT and Iris and PixInsights (and GIMP) all have a different view of the histogram

 

Hopefully I will get a clear night and can go and try this. I have had horrid vignetting on my previous images and I am really hoping this eliminates / mitigates at least one source of that

 

I like that this is controlled and scientific 


Edited by limeyx, 25 September 2020 - 11:06 PM.


#7 russellmm

russellmm

    Vostok 1

  • -----
  • Posts: 151
  • Joined: 02 Aug 2020
  • Loc: Gig Harbor WA

Posted 25 September 2020 - 11:30 PM

remember that if you have Pixinsight setup to import as "Pure Raw" they are going to come in there as grayscale images. So when you stretch it, you do not see it as a blue image but rather as gray. Your flats are looking pretty much just like mine now. They should work really well.

 

This is what it looks like in Pixinsight as a pure raw grayscale superstretched.

 

Attached Thumbnails

  • 2020-09-25 21_29_47-PixInsight.jpg

  • limeyx likes this

#8 limeyx

limeyx

    Ranger 4

  • -----
  • topic starter
  • Posts: 355
  • Joined: 23 May 2020
  • Loc: Seattle, WA

Posted 25 September 2020 - 11:32 PM

remember that if you have Pixinsight setup to import as "Pure Raw" they are going to come in there as grayscale images. So when you stretch it, you do not see it as a blue image but rather as gray. Your flats are looking pretty much just like mine now. They should work really well.

 

This is what it looks like in Pixinsight as a pure raw grayscale superstretched.

Thank you, that makes total sense as I do have pure raw checked. I am hoping to put the flats issue behind me and this really helps, I really appreciate your response



#9 endlessky

endlessky

    Viking 1

  • -----
  • Posts: 556
  • Joined: 24 May 2020
  • Loc: Padova, Italy

Posted 26 September 2020 - 02:08 AM

I used to struggle a lot with - DSLR - flats, too.

 

Until I discovered these:

 

- the half peak histogram as viewed from the back of the camera doesn't work. It has to be half of the total ADU, when viewed in a linear preview (EKOS, PixInsight, ecc.). The max ADU is 16384, so I aim for a linear mean/meadian histogram reading of 8000

 

- the flats - and the lights - need to be both calibrated at least with bias (I don't use darks). If I don't calibrate them with bias, the flats would either overcorrect or undercorrect: the dust spots would either stay dark or turn lighter than the sorrounding background

 

After I started doing the above, my flats finally started working perfectly and I never had any more problems of under - or over - corrected vignetting and dust spots.


  • limeyx likes this

#10 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 6,358
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 26 September 2020 - 04:14 AM

I think i do the same, i use flats, and bias, no darks. Allthough the latter one is debatable, especially now i starting to get longer subs.

 

Do they work , well they get rid of the dust spots , and i do have dust spots for sure. Many bad things come up while processing but no dust spots sofar...fortunately


  • endlessky likes this

#11 endlessky

endlessky

    Viking 1

  • -----
  • Posts: 556
  • Joined: 24 May 2020
  • Loc: Padova, Italy

Posted 26 September 2020 - 04:22 AM

The only reason I don't take darks is because I am using an uncooled DSLR and I have no way of making sure I am matching the temperature of the lights with the temperature of the darks.

 

Other than that, the D5300 has fairly low noise and once I made sure I swamped the read noise enough and dithered by a good amount, noise is quite easy to deal with in post-production.


  • F.Meiresonne likes this

#12 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 6,358
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 26 September 2020 - 04:34 AM

I used to take them , but after a discussion with 'The Elf' and regarding that then did only 60 sec subs it seemed not worthwhile for my Canon 800d.

 

Think maybe i will start making some, but i do this after i have taken my lights, thesame evening, i think temp should be pretty ok then.

But  it takes extra time to spent...on a working day not evident...


  • endlessky likes this

#13 endlessky

endlessky

    Viking 1

  • -----
  • Posts: 556
  • Joined: 24 May 2020
  • Loc: Padova, Italy

Posted 26 September 2020 - 04:59 AM

Time is definitely an issue, if you have a DSLR. Taking them the same night means cutting into your imaging time.

I think one of the obvious reasons to get a dedicated camera, beside the huge noise decrease at 20-30° below ambient, is that you can easily match temperatures even in the daytime, indoors, while the camera is plugged in the PC. Perfect temperatures and no imaging time wasted!

#14 limeyx

limeyx

    Ranger 4

  • -----
  • topic starter
  • Posts: 355
  • Joined: 23 May 2020
  • Loc: Seattle, WA

Posted 26 September 2020 - 03:00 PM

Thanks everyone. Hopefully I am back on track now. It is supposed to be clear tomorrow so will see.

 

I have been taking darks but as pointed out, they do eat into imaging time so I may stop doing that ...I dont have enough data yet but if I may take darks tomorrow and then process both with and without darks and see what's what




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics